12dw7 for cascode?

[Austin]

Seems like the problem with cascode of the popular 12ax7 is the output impedance is high and the bottom tube seems like would have a hard time being effective at controling the low impedance of the cathode of the upper half. What about a 12dw7 then? The 12au7 section on bottom could would be more effective I think, the best of both worlds. A 12au7 has small gain and the 12ax7 has small current but used together they might be a good match. Most of the gain comes from the top tube so put the 12ax7 up there and put the 12au7 side on bottom where its current could be better used. I tried googling this, it seems like it would be a good idea but didn't get many results.

I was reading another post about a failed cascode pedal attempt, and that got me thinking.. I have one of these in my box o tubes.. Maybe I could try to build my own booster pedal with one.

[Gingertube]

Austin,
The lower device enhanced current arrangement of the cascode (additional resistor from B+ to lower triode anode) would be the way to go to get the "12AU7'ish" triode current up a bit for more gm.
This method is described in Merlin's Preamp book. Remember that the gain of a cascode is approximately the lower device's gm x the upper device's RL.
You would also want to elevate the heater to say +70V.

Definitely worth a try.

Cheers,
Ian

[Austin]

Austin,
The lower device enhanced current arrangement of the cascode (additional resistor from B+ to lower triode anode) would be the way to go to get the "12AU7'ish" triode current up a bit for more gm.
This method is described in Merlin's Preamp book. Remember that the gain of a cascode is approximately the lower device's gm x the upper device's RL.
You would also want to elevate the heater to say +70V.

Definitely worth a try.

Cheers,
Ian

That's all the encouragement I need thanks.. I feel fascinated by cascodes lately and it's good I have a place like this to ask my questions to. Thank you.

[Gingertube]

Austin look here for the basic cascode calcs.
The Valve Wizard

If you use a 12DW7 then the "standard" circuit will be runnning too low a current. Current will be limited to about 1mA by teh 12Ax7'ish top triode. That is too low a current for the 12AU7'ish bottom triode and tubes run at low current exhibit low gm. So you need to give the bottom triode a bit more current - get it up to at least 3 otr 4 mA although more would be better. The way to do that is to add another resistor from B+ to the anode of the bottom tube, say 100K. That feeds extra current around the top triode directly into the bottom triode (that is, the extra current bypasses the top triode). If you can scrape together a few bucks I thoroughly recommend both of Merlin Blencowe's books which you can see on his Valve Wizard website. I have both. The preamp book in particular addresses a surprising number of the issues which get raised here.

Cheers,
Ian

[Amp Kat]

Cascodes are used lots in Mic-Preamp applications and used for current sourcing or bootstapping circuits. Check out this link it's a jensen tube mic-pre and I've built it and it sounds very good. They are very linear circuits and very clean.

http://www.jensen-transformers.com/as/as054.pdf

[tubeswell]

Interesting that the plate resistance of the 12AU7 triodes are used to form the load in that example that Amp Kat posted. The upper one would want about 60V heater elevation I guess. (What elevation have you got that heater on Kerry, if any? - Just out of curiosity)

[Amp Kat]

Interesting that the plate resistance of the 12AU7 triodes are used to form the load in that example that Amp Kat posted. The upper one would want about 60V heater elevation I guess. (What elevation have you got that heater on Kerry, if any? - Just out of curiosity)

Yes you are spot on and I used about 100 volts and it's night and day from either 6.3 or 12 volts dc. Jensen says to use very clean power but they never mentioned elevation but it's almost a must and it cleans it up nicely. Lots of filtering in the plate supply is also better and a beefy transformer sounds much better than a small one like the Hammond 269 series. I used a MOSFET regulator and an old Marshall power tranny and tweaked it down and it sounded the best. Very big and punchy. The input tranny and output are also key elements so not sure how this would work out for Guitar as I haven't tried it but would think it would be very clean and would work nice into an interface for computer recording. I'm sure a nice Toroidal like the Van Der Veen one from Plitron would sound awesome too.

[Austin]

So what happens if I just bias the 12ax7 to allow say, 4 milliamps? Shorter tube life? How much shorter? It might be acceptable if it sounded really good...

[tubeswell]

In terms of the 12AX7 plate current, seeing as how a 12AX7 is rated for 1W, that would mean running 4mA of plate current would require not exceeding 250V on the plate. Shorter tube life? The sign will be when it starts to red-plate :-) JJ optimistically rates the maximum plate current of their 12AX7 at 8mA (but doesn't say whether this is for one triode or not)

From what I can gather, in a cascode the bottom triode controls the current flowing into the upper triode, so the gain of the stage is dependent on the transconductance of the lower triode. If the lower triode has a higher transconductance, the stage will have more gain. So a 12AU7 would be a better lower triode than a 12AX7 in this regard.

Also if you increase the current through the whole cascode the upper triode needs to be biased hotter, which gets closer to 'screen grid' current and limited headroom.

As an aside (from what I can gather), ECC88s make the 'best' dual triode cascodes, and 12AU7s are a close second, in terms of the achievable output swing for a given HT voltage. But I've never tried either of them.

I'd be interested in the results of any experiment you may care to undertake.

[Austin]

In terms of the 12AX7 plate current, seeing as how a 12AX7 is rated for 1W, that would mean running 4mA of plate current would require not exceeding 250V on the plate. Shorter tube life? The sign will be when it starts to red-plate :-) JJ optimistically rates the maximum plate current of their 12AX7 at 8mA (but doesn't say whether this is for one triode or not)

From what I can gather, in a cascode the bottom triode controls the current flowing into the upper triode, so the gain of the stage is dependent on the transconductance of the lower triode. If the lower triode has a higher transconductance, the stage will have more gain. So a 12AU7 would be a better lower triode than a 12AX7 in this regard.

Also if you increase the current through the whole cascode the upper triode needs to be biased hotter, which gets closer to 'screen grid' current and limited headroom.

As an aside (from what I can gather), ECC88s make the 'best' dual triode cascodes, and 12AU7s are a close second, in terms of the achievable output swing for a given HT voltage. But I've never tried either of them.

I'd be interested in the results of any experiment you may care to undertake.

Thanks for the reply Tubeswell. Since the 12ax7 top half would have it's grid fixed to a reference point and the cathode moves, it is certainly possible to opperate in class a2 with grid current if the cathode were to dip down too far- part of the reason for my fascination with the cascode. I don't see this as a problem however or that it would limit the headroom because the 12au7 bottom part would be capable of keeping up with the current demands of the top. This assumes a bias that can't change though, so a resistive voltage divider won't work. Dual positive supply voltage with middle voltage as a reference, string of zeners, or buncha nine volt batteries from ground are some ways I can think of. I heard zeners are noisy, but I don't see how they would make any noise since they would be conducting dc and there would be no switching. Is that right?

[tubeswell]

The upper triode could be fixed-bias with a bypass cap on the lower leg of the divider to keep the g2 voltage stable. The bypass cap would want to be big enough to keep the grid voltage steady down to the the lowest desired frequency. (Or you could use a zener on the lower leg with a bypass cap to snub switching noise).

You could cathode-bias the upper triode with an Rk||Ck between the lower plate and the upper cathode as well as a Rg resistor (say 1M) from the plate of the lower triode to grid of the upper triode. You would still have bypass cap from the upper grid to ground to keep the upper grid voltage steady.

Or you could use grid-leak-bias of the upper triode, where the lower plate and the upper cathode are directly connected and you have a largish (say 1M minimum) resistor from that junction to the upper grid. Again you may want a bypass cap from the upper grid to ground. With this option, you could have the current boosting that gingertube talked about before (with a resistor from the B+ to the plate/cathode junction) to supply the lower triode with more current (i.e. an 'extra' plate resistor for the lower triode). This resistor needs to be sized so that the voltage drop across it = B+ - the lower plate's voltage. See pp 154-155 of Merlin Blencowe's book to work it out. This would probably be the ticket if you used a 12AU7 triode for the lower triode and a 12AX7 triode for the upper triode (but I'm only speculating about this. Others here would know a lot more than me about this.)

[Austin]

Witha dual positive supply why not tie the upper tube grid to the middle rail and the top tube plate to the top rail? The top tube cathode voltage would be controllable with the bias of the bottom tube and the grid could be adjusted a with a 9 volt battery if it needed. Wouldn't this be the easiest?

[tubeswell]

Hi Austin

Do you have a schematic for what you have in mind (that you can post)?

[Austin]

Hi Austin

Do you have a schematic for what you have in mind (that you can post)?

I drew this with TinyCad to give the basic idea, in a pedal it would probably use a 200 volt supply so as to keep to the heater to cathode limit. This is drawn with the power supply I just made as a reference and I will use a wall wart to power the heater.

[tubeswell]

Looks interesting. Maybe you should also consider a grid stopper resistor for the upper triode.